Effects of DF are usually investigated using urine, blood and faecal analysis, but exhaled breath is of increasing interest because it allows non-invasive sampling. There are almost 900 VOCs in the exhaled breath of healthy humans (16) which could be considered as potential biomarkers of metabolic disorders, specific diets, or both. Consistently, breath analysis (mainly hydrogen and methane) is already used in clinics for monitoring gastrointestinal disorders (12, 17–19). Recently, Rondanelli et al. reviewed biomarkers present in exhaled breath of gastrointestinal diseases and nutritional status (20). For example, exhaled acetone has been reported to be an important biomarker of diabetes (9). A few studies have also monitored the effects of diet (gluten-free diet, high versus low fat dairy drink, high versus low fiber diets) on the composition of exhaled breath (12, 21–24). In the present study, we compared for the first time the effect of a unique ingestion of an insoluble DF (CG prone to be fermented by the gut microbiota) versus maltodextrin (a fully digestible carbohydrate) as placebo on BVM profile in young healthy adults. The interpretation of our data considered characterization prior intervention in terms of faecal microbiota composition and faecal SCFA. We hypothesized that some BVM may be the reflect of DF gut fermentation with respect to inter-subject variation.
It has already been described that the presence of some metabolites such as acetone, ethanol, isoprene and methanol in the breath is conditioned by the fasting state (11). In our study, the highest levels of acetone but also acetonitrile, acetic acid, propionic acid, caproic acid, 1- and 2-propanol, benzene and benzaldehyde were found at fasting state in the morning. The potential origin of such BVM was already discussed elsewhere focusing on diet, environmental exposure (exposure to air pollutants) and biological pathways (11, 25). Some of them were decreased after the breakfast without any further variation due to the lunch, the collation or the CG intake whatever the time considered (1- and 2-propanol, benzene and benzaldehyde). We hypothesized that the presence in breath at time 0 h of this kind of BVM with high affinity with fat tissue may be the result of the high rate of lipolysis at fasted state. Other BVM were different between the pre- and postprandial states without any significant influence of CG intake. It is the case of acetic acid, (iso)valeric acids, acetone, acetonitrile, ethanol, isoprene, methane and H2S for which levels increased after the lunch in both groups. Acetone, produced via fatty acid oxidation, has been one of the earliest identified gaseous disease biomarkers, being distinctively present in the breath of patients with diabetic ketoacidosis, hypoinsulinemic states, and starvation (26). Subtle acetone fluctuations also systematically occur in response to modest, physiological changes in insulin/glucose levels (26). In contrast, ethanol is not produced by human cells, but small concentrations in blood and breath derive from alcoholic fermentation of glucose by gut bacteria expressing pyruvate decarboxylase. Following carbohydrate ingestion, blood and breath ethanol increase (26). Furthermore, the levels of exhaled breath ethanol increased after intravenously administered glucose load, supporting the hypothesis that glucose enrichment of gastrointestinal capillaries per se may stimulate gut bacterial fermentation to some degree (27). The source and physiological effect of isoprene in humans are matters of debate. In animals and humans, this intermediary metabolite of cholesterol synthesis is formed in the liver from isopentenyl pyrophosphate (IPP) and its isomer (DMAPP) (28). However, as this reaction is slow and may be insignificant at physiological pH values, it is unlikely to completely explain the endogenous isoprene production by animals. So far, two major metabolic pathways leading to DMAPP have been identified: the mevalonic acid (MVA) pathway and the 1-deoxy-D-xylulose-4-phosphate/2-C-methylerythriol 5-phosphate (DOXP/MEP) pathway. The DOXP/MEP pathway was demonstrated to prevail in plants and most bacteria, whereas the MVA pathway is mainly present in higher eukaryotes. In bacteria DMAPP is converted into isoprene enzymatically by isoprene synthase. Surprisingly, isoprene increased with the time of fasting and fell down in postprandial state independently of the DF intake. Another study showed that isoprene concentrations did not change significantly following feeding with a liquid protein-energy meal (29). It would be interesting to assess the role of isoprene as marker of cholesterol synthesis, that can be affected differently following meal composition. A pilot study of 7 healthy subjects suggested that increased metabolic activity in the gastrointestinal tract after ingestion of a meal could explain, at least partly, the increase of BVM such as ethanol, 1-propanol, 3-hydroxybutanone, propionic acid, and butyric acid (12). Here, we did not observed changes with meal of propionic acid and 1-propanol. Anyway, our data led us to conclude that a lot of BVM from our targeted analysis, principally produced by host metabolism, reflected metabolic state rather than a response to a DF ingestion. In addition, our study revealed that bacterial metabolites coming from the gut microbiota and exhaled in the breath such as H2S, methane and SCFA (acetic acid, butyric acid, (iso)valeric acids, propionic acid, caproic acid) may be influenced by the ingestion of a meal or by the fasting state.
Hexose and pentose sugars are fermented by isolated human colonic bacteria via pathways leading to the formation of SCFA, ethanol and hydrogen depending on the strain and species (30). In particular, butyrate formation occurs in Firmicutes bacteria, either via butyrate kinase (in many Clostridium and Coprococcus species) or via butyryl CoA:acetate CoA transferase (31). Marzorati et al have shown that fermentation of CG led to an increased production of both propionate and butyrate in vitro (5). In the present study, a single intake of 4.5 g CG led to an increase of exhaled butyrate suggesting that this SCFA is not restricted to enterocyte metabolism, but may be absorbed in the circulation and then released in breath as a signature of gut fermentation of CG, since the increase did not appear when volunteers received placebo. We were unable to relate exhaled butyrate to faecal butyrate. It should be kept in mind that SCFA production mainly occurs in the proximal part of the colon where the availability of substrates is most abundant. Accordingly, up to 95% SCFA are rapidly absorbed by the colonocytes resulting in decreasing concentrations from the proximal to distal colon and only about 5% is excreted in faeces (32). In addition, intestinal microbial fermentation is a dynamic process influenced by a wide range of factors. Therefore, the levels of each metabolite are a result of metabolic fluxes of highly variable rates, which are not adequately represented in steady-state metabolite profiles in faeces. Our kinetic analysis offered the possibility to assess the dynamic processes of gut fermentation. Interestingly, exhaled butyrate was positively correlated with Mitsuokella from the Firmicutes phylum. It has been demonstrated that concentration of exhaled methanol increased from a physiological level of ∼0.4 ppm up to ∼2 ppm a few hours after eating around 500 g of fruits, the effect being principally due to pectin fermentation (11, 33). Here, the basal levels of breath methanol were in the same range than the physiological level but did not reach a concentration higher than 0.5 ppm. In addition, bacterial fermentation in the mouth or throat can partly explain increased postprandial levels of exhaled breath BVM (12). Breath methanol increased already just after taking the CG with the breakfast, suggesting that this metabolite was not a reflect of CG fermentation in the lower part of the gut. The kinetics of triethylamine release was different, since it increased 4 h after CG intake with a peak obtained at time 6 h as observed for butyric acid but the range of changes was very weak (maximum 0.005 ppm). Although no literature exists about metabolic pathway explaining a potential origin from gut bacteria, the kinetic study performed in the present study supported the hypothesis that this BVM appeared following gut fermentation of CG as observed for butyric acid. We found that production in the breath was positively correlated with the presence of Mitsuokella and Catenibacterium in the faecal matter. Breath ethane decreased after CG intake whereas pentane followed the same kinetics than exhaled butyric acid. The presence of pentane in exhaled breath is considered as a result of lipid peroxidation of polyunsaturated fatty acids in cellular membranes, a process mediated by free radicals and oxidative stress (10). In contrast to butyrate, those considerations do not allow to state that the increases of pentane and methanol due to CG ingestion come mostly from direct bacterial metabolism. Data show that lactobacilli can produce 3-hydroxybutanone (acetoin) and 2,3-butanedione (diacetyl) (34). The production of 3-hydroxybutanone and 2,3-butanedione has already been studied on different Lactobacillus strains growing on Jerusalem artichoke juice rich in inulin that served as substrate (34). Their production is related to citrate metabolism through the citrate–oxaloacetate–pyruvate–acetolactate–acetoin/diacetyl pathway, where pyruvate is considered as the precursor. In our study, although 3-hydroxybutanone and 2,3-butanedione were exhaled after CG ingestion, no correlation was found with lactobacilli. Interestingly, 3-hydroxybutanone production in the breath was positively correlated with Mitsuokella, the same bacteria revealed for butyric acid production. The genus Mitsuokella, comprising only two named species, are Gram-negative obligate anaerobes which utilize fermentable carbohydrates to form acetate, lactate and succinate (35). Of these two species, M. multacida is regularly found in the intestinal tract of humans and appears to be responsive to diet changes. The acids produced were potentially available to other acid-utilising bacteria for the formation of butyrate, such as Roseburia (36).
In conclusion, this study showed that CG intake changed the exhaled BVM profile and that meal intake influenced the BVM profile. Even if the limitation of the study is to focus on a small cohort and refer to hydrogen producers only, we have validated a protocol allowing to analyse the effect of DF intake in the morning on the profiling of breath metabolites throughout the day, showing that post prandial - post absorptive period is probably the most relevant timing for the evaluation of BVM. Exhaled butyric acid, triethylamine, 3-hydroxybutanone and 2,3-butanedione are reflecting the bacterial metabolism of CG. This implies that analysis of BVM in exhaled air can be used as a dynamic approach to noninvasively assess the responses to DF intake related to gut fermentation.